MATHEMATICAL INTERPRETATION OF ASSOCIATIVE MODELLING PROBLEMS


Dr. An. Matvejevs 1 and G. Vulfs 2
1 Probability Theory & Mathematical Statistics Dept., Riga Technical University, Meža iela Ľ, Riga LV-1048
E-mail: anmatv@egle.cs.rtu.lv
2 Information Technologies Institute, Riga Technical University, Meža iela 1/3, Riga LV-1048
E-mail: vulfs@itl.rtu.lv

ABSTRACT:The proposed associative algorithm approach to dynamic systems is universal, and it allows one to solve applied problems with only scant initial information about the object in protocol form. The associative method is based on the use of precedent situations fixed in the protocol of observations. This allows comparison of the associative algorithm with the method of local regression. Both methods are nonparametric, and they do not require a priori information on the structure of the dependencies synthesized.

KEYWORDS: Associative modelling algorithm, local regression

---

Back to HMS2003 Home Page